DESIGNING AND ASSESSING THE BENETITS OF CONTROL POLICIES FOR CONVEYOR 192
BELT SYSTEMS IN UNDERGROUND COAL MINES
by
Dr.£,F, Wolstenholme
. B,Se., PHD., M.Tech,,
C.Eng.y
University of Bradford,
ABSTRACT
This paper describes the application of System Dynamics in what is
traditionally a hard engineering area, but where the application of
analytical techniques are limited by the stochastic nature of the system
driving forces (coalface output rates) and the need for highly credible,
uanagezent orientated results, Methods of analysis have thus centred
on using discrete siculation techniques based on open system models, primarily
to assist in capacity design.
The use of System Dynamics in this context is based on two premises, The
first of these is that System Dynamics has, in addition to its softer areas
of application, considerable potential to both supplement and enhance
Operational Research approaches in the analysis of such systems, Secondly,
it is the author's belief that the key to further development and acceptance
ef System Dynamics lies in bridging the gap between itself and associated subject
fields, such as Operational Research, by direct demonstrations of the approach
within these fields, Recent technological advances within the coal clearance
field have provided an excellent form for such a demonstration.
The trend in the installation of micro computers for centralised monitoring
of the state of underground conveyor belts and bunkers, is rapidly increasing
and the scope for totally automatic, real time control of these systems has,
consequently, been greatly enhanced. However, the progress in information
retrieval has outstripped the development of compatible advences in methods
ing total system controls to make best use of collected information,
and the research work described here concerns how System Dynamics can assist
such design. This centres on the development of a System Dynamics model of
an underground conveyor belt syste, incorporating realistic production
generated patterns, The model is used to test out and improve the design of
alternative policies for bunker discharge rates under a wide range of system
parazeters: The findl policy evolved in this way has general applicability
and is shown to maximise conveyor belt utilisation, be independent of limitations
in the saxirum bunker discharge rates and to require only the monitoring of
bunker levels. Finally, the model is used to quantify the benefits of such
inproved control in terns of savings in physical capacity required to obtain
naximum system efficiency.
1.
INTRODUCTION
The majority of Coal Mines utilise conveyor belt systems to
transport coal from the working coal faces to the surface. The major
problems in designing the capacity of such systems are that coal face
output rates fluctuate over time, due to variations in the shift patterns
in operation, variations in the coal cutting speed (due to geological
changes) and the reliability of coal face machinery. Further, since
reserves of coal are depleted by extraction, the layout of coalfaces in
a colliery is a geographically dynamic phenomena.
‘The capacity of coal clearance systems must hence be designed to
cater both for short term fluctuations and longer term changes in coal
-production rates, As # result of major developments in coal mining technology
over recent years, and the consequent trend tovards concentrating production
on fever and Larger coalfaces“, the situation vhere colliery coal
clearance systems are working at or above their design capacity is
frequently encountered, Consequently, attention is now being focussed
on the use of more sophisticated control of these systens in order to make
the best use of existing capacity, The feasibility of such control has
been enhanced of late by technological advances in the development of
mini and micro computers for real time control.
the installation of small computers is currently taking place on an
increasing scale at colliery level, and these are being used to monitor
and display up to date information (on both coal clearance and many other
underground systems) in central control stations, The implenentation of
action based of this information is at present largely manual ©) (3)(4)(5) (6)
but the ultimate potential of these applications is that decision rules or
control policies can be automated and, hence, control action can be fed
directly back to the operations,
‘The fundamental difficulty in attaining this potential in any
information and control system are those of determining which information
sources to monitor, and what form of control rule to use, This presents
souewhat of a dilemma, because control rules cannot be formulated unless
@ choice of information has been made, and it is difficult to choose which
information to monitor unless the benefits of using it in control rules
have been assessed,
2
System Dynamics is a technique for applying control engineering ideas
to complex management systems”) and system dynamics models have been used
in a wide range of industrial and socio-economic systems, to help overcome the
apove mentioned probe The technique involves modelling a system in terms of
its constituent Levels and rates and developing this into a continuous
simulation model, incorporating both physical flows and information feedback,
The merits of alternative forms of rate (control) equations can, therefore,
be investigated based on a variety of information inputs,
To achieve this end continuous simulation software is available and
the main purpose of the current research has been to investigate the merite
of applying such a general suite of progranmes, (DYSMAP)."®2 to study
the specific issues of coal elearance, This approach is intended to
provide an interesting contrast with the well developed and long standing
discrete simulation programmes currently used in the British Coal Industry
in this context, These programmes have been extensively and successfully
used to study problems of coal clearance capacity, but only lately have
been extended to allow investigations of control, F
MODEL DESCRIPTION
In order to test the general feasibility of building continuous
sizulation models, capable of investigating alternative aspects of control,
a rodel has been developed of a simple three bunker coal clearance
system, It is assumed that each bunker is fed by a single coalface and
that the bunkers discharge ontoa drift conveyor, which transports the coal
directly to the surface of the mine, The physical flows of this eysten are
shown diagramatically in Figure 1, The overall aim was to develop and test
alternative discharge policies for the bunkers, measured against an
efficiency criterbn based on the ratio of cumulative coal output cleared to
the surface, over a given period of time, to that potentially available for
the three coalfaces, The model vas developed in moduler form so that at
@ later stage any configuration of faces, belts and bunkers could be represented.
Figure 2 shows a slightly more detailed, but still simplified,
diagram of the model, indicating some of the composite steps involved in
building up the sectors for coalface 1 and bunker 1, The diagram
includes some information feedback links to demonstrate how bunker discharge
policies were incorporated in the model, This structural representation,
which vas replicated for the other coalfaces and bunkers, is explained in
the following sectione,
3.
In order to produce a realistic pattern of coal output over tine, the
coalface sectors of the model were designed to take into account shift
working times, variations in the rate of coal cutting and stoppages due to
machinery breakdowns, Figure 2 shows how the first two of these factors
are initially superimposed onto a "base' coalface output rate to model
the pattern of available coal output over time, It was assumed that
once randomly set the available coal output rate would hold for 30 minutes
before being reset, The actual coalface output is then generated by
superimposing randomly generated coalface breakdowns, This is achieved by
alternatively sampling a length of production run and a length of breakdown
period from normal distributions for these factors, with adjustable mean
and sténdard deviation,
Full details, including equations, of the coalface model used
to generate the output will be found in a separate paper (9).
Examples of the dynamics produced by the model, at each stage of the
procedure, are presented in Figure 3,
The actual coalface output rate is then fed directly into the bunker (see
with a provision that the coalface will be switched off if the bunker Fig.2)
capacity is exceeded, During such periods the cumulative available output
rate is measured to represent the lost production due to inadequate bunker
capacity,
The bunker itself is discharged onto the conveyor belt, according to
the discharge policy to be investigated, For dempnstration purposes :
Figure 2 gives an example of how such a discharge policy might be constructed,
Here, it is assumed that the desired rate of discharge of the bunker will be
‘a function of the bunker level (or other bunker levels), and that the actual
discharge rate will follow this, subject to there being coal available in
the bunker and space available on the conveyor belt at the discharge point.
The actual discharge rate, once determined, then depletes the bunker level
and the coal leaving the discharge point is the cum of that arriving
and this discharge rate, Additionally, the coalface is also svitched off
whenever the desired discharge rate exceeds the belt capacity available and
the bunker is full, Cumulative coal lost during such periods is measured
and classified as losses due to lack of belt available capacity.
Ay
Figure 2 also shows how the cumutative coal reaching the surface
is compared with that available at the coalface, to produce the efficiency
measures
By converting Figure 2, and similar representations of the other
coalfaces and buskers, into explicit equations representing the influences
shown, a composite simulation model is produced, This can then be run over
many time periods to investigate the overall performance of the systen under
different patterns of coalface production, aifferent values of coal
‘clearance system paraneters and different bunker discharge policies.
Coalface 1 Coalface 2 Coalface 3
wea
Figure 1
GENERAL LAYOUT OF SITUATION MODELLED
shift times
“4
\.
——" bet
Actual Face 1
Generation rate ~
|
ven > -
Output rate
*
Available Coalface 1
\.
ey
” pase Coalface 1
‘Output rate
INFLUENCE DIAGRAM FOR COALFACE 1. AND BUNKER 1
Random Breakdown.
i
in Coal Cutting rate
Random Variat:
FIGURE 2
Output rates
Coalfaces 2 & 3
overall
Efficiency
Available
from
° oi
3 Bo
FF di
4a ve 3)
i we
ri Z
mat a a
%
ia
a 2&8
N 3 [ |
Pe
ie aye 34
3 Bag, See
B\\\ #BOH
B88
u 3
oy
a8 |
; Na ag
ood
iN
ING
ag
:
3
4
2
as
#
u
Bou
Na
‘ Pet
Nu
aa
35
ica,
gE
(2H/st01) eavy andang pdsymopuly
6:
P g
or oS i”
3 g fe 2
3 a 8a g
2 & a2 0C«s gs
g “¢ 88 ay
é ig P
4 a.
s .8 2
fa ‘#3
. 4
ee 3
: = 5
.& . y
t T Tv | iS
8" § 8 € 8
(AH/suol) eary andang aqqetreay
fog g 2g
1000.4
i
22504)
gegdgeig:
(2H/su03) 2aey andgng penqoy
SHIFT
SHIFT
GRAM,
FIGURE 3 DYNAMICS OF COALPACE oUTPU
195
ny
EFFECTS OF ALTERNATIVE BUNKER
_ DISCHARGE POLICIES
Design of Experinents
ALL experiments and results presented in the report are based on varying
the coal clearance paraneters and bunker discharge policies, Connon features
of all the model runs were that the conveyor belt could operate throughout the
working day, but that the coalfaces vere operated over two six hour shifts,
each with a base output of 1000 tons per hour and a range from 600 to 1400 tons
per hour, The length of production and breakdown periods were. specified
separately for each coalface but mean lengthswere in the order of 120 minutes
and 10 minutes respectively, Also the model was run in each case for one
complete day (24 hours) starting from an equilibrium situation and using the sane
stream of fandom numbers for each coalface on each run.
Fach experimental model run was defined in terms of the coal clearance
parameters used, and these runs are Listed on the left hand side of Figure 3,
against which various bunker discharge policies were subsequently tested, Although
the three bunker coal clearance system model outlined represents a simple type of
such a system, it can be used, by suitable choice of parameters, to model a vide
range of conditions encountered in practice, The experiments defined in
Figure 3 can be seen to fall into two blocks,
In the first block the belt capacity (2000 tons/hour) is chosen to be less
than the sum of the base output rates of the three coalfaces, and hence these
parameters represent a storage bunker situation, Here, the purpose of the bunkers
is to act as storage over the day to match face outputs with inadequate belt
capacity, and scope exists to trade off bunker capacity against improved control ~
of bunker discharge rates, Four experiments are defined here which examine the
effects of lowering the maximum bunker discharge rates from their base value of
1000 ton/hour to 700 ton/hour, and increasing the capacity of all bunkers from
their base value of 500 tons to 1000 tons and 1200 tons respectively,
‘The second block of experiments represent a surge bunker situation, where the
belt capacity is first set to 2,500 tons/hour which (after allowance is made for
coalface breakdowns) is approximately capable of dealing with the sum of the base
output rates of the three coalfaces, but not the sum of their maximum rates.
Here, the purpose of the bunkers is to smooth out instantaneous flpctuations in
coalface output rate over and above belt capacity. Obviously less bunker
capacity is required in these circumstances, than in the storage bunker situation,
and experiments are hence defined in terms of 150 tons and 500 tons bunker capacity
respectively, ‘The belt capacity is secondly set to 3500 tons/hour, which is
considerably more capable of dealing with instantaneous peaks in coalface outputs,
and similar bunker capacity increases examined,
&
Simple Bunker Discharge Policies - definition and results 196
The model was first tested using rather obvious and crude discharge
policies as follows:
Policy I: Fixed Discharge Policy
Here it was assumed that each bunker could only be operated at zero or
its maxirum discharge rate, and that the latter would be used as long as there
vas coal available in the bunker and room available on the belt,
Policy IJ t Variable Discharge Policy
Eere it was assumed that each bunker discharge rate could be set at any
point between zero and the maximum discharge rate and would be set in proportion
to the bunker level:
Bunker level * Maximum Discharge Rate «@
Bunker Capacity
i.e, Desired Discharge Rate =
Again subject to coal being available in the bunker and room being available on
the conveyor belt,
The result of applying these simple bunker discharge policies
to the runs previously defined are shown in Figure 4,
DEFINITION OF EXPERIMENT OVERALL SYSTEM EFFICIENCY
Total | Maxinum Capacity Bunker Dischar; i
a ge | Bunker Dischai
Belt | Discharge | of each Policy I Policy I
Capa~ | Rate of Bunker a ‘
city each. Bunker i (Fixed Discharge) | (Variable Dischar
l(tons/ | tons/Hour | (Tons) a
Hour)
BLOCK 1/Pun 1} 2000 1000 “500 70.1 719
STORAGE /Run 2] 2000 700 500 53.7 69.1
Fun 3} 2000 1000 1000 72.9 76.7
SYSTE |Run 4{ 2000 1000 1200 73.9 787
BLOCK 2] Run 5} 2500 1000 150 654° 75.8
SURGE | Run 6] 2500 1000 500 70,1 84.0
BUNKER [Run 7] 3500 2000 150 83,7 714
SYSTEM | Run 8) 3500 1000 500 91.8 85,8
Figure 4 Overall System Efficiences for
x
With a conveyor belt capacity of only 2000 tons/hour (block 1 results)
and a maximum discharge rate from each bunker of 1000 tons/hour, Policy I
exhibits efficiences between 70.1% and 73.9% (runs 1, 3 and 4), since, at
best only two bunkers can discharge at any one time. The difference.
between these runs demonstrates the increased efficiency resulting from
increasing the bunker capacity and run 2 indicates the reduced efficiency
resulting from a lowering of the maximum bunker discharge rates. When the
conveyor belt capacity is increased to 2500 tons/hour (runs 5 and 6 in block 2)
no imptovement in efficiency results because, due to the policy, it is still
only possible to discharge at most two bunkers at any time. Consequently,
run 6 gives the same result as run 1 and run 5 is reduced to 65.4% due to
less bunker capacity available. Increasing the belt capacity to 3500
tons/hour results in improved efficiencies (runs 7 and 8), since all three
bunkers can now be discharged together. (This result would also be
expected at a conveyor belt capacity of 3000 tons/hour).
Under policy I the discharge rates from the bunkers can only be either
zero ot the maximum discharge rate. Hence there are times, particularly
between shifts, when some discharge could take place but doesn't (because
this would be at less than the maximum rate), and it is to be expected that
: the conveyor belt capacity is somewhat under utilised. One of the most
important conclusions from the experiments in this section is that wher both
sufficient bunker and belt capacity exist the maximum theoretical efficiency
for the system (given the coal face breakdown pattern used) is attained (run 8).
This result gives confirmation of the fact that if total system capacity is
adequate then the only control necessary over bunker discharge rates, is to
deploy their maximun setting.
The results from using Policy II are also shown in Figure 4. This
variable bunker discharge policy overcomes the major deficiency of the
previous policy, in that discharge can now take place at any internediate
Fate between zero and maximum, As a consequence improvements in overall
efficiency are achieved in most cases relative to Policy I. This is
particularly noticeable at a conveyor belt capacity of 2500 tons/hour,
where, unlike Policy I, advantage can now be taken of the additional 500
tons/hour conveyor capacity available over block 1 experiments. The
exceptions to the improvements in efficiency, relative to Policy I, occur
in runs 7 and 8, associated with a conveyor belt capacity of 3500 tons/hour.
9a.
These reduced efficiencies do in fact highlight one major veakness of
the policy used, Once sufficient conveyor belt capacity exists the
bunker discharge rates under Policy II are simply a function of the bunker
level (see equation (i)) and the maximum discharge rate is only employed
when the bunker is full. Consequently, if the bunker capacity is almost
adequate to cope with the coalface output (run 7), the bunkers are rarely
full and the discharge rates attained are less than the maximum. As the
bunker capacity is increased the discharge rate associated with a given
bunker level decreases and the maximum theoretical efficiency for the
system is never attained (run 8). This is an interesting disadvantage of
vhat would intuitively appear to be a sound discharge policy. It should,
however, be noted that at lower conveyor belt and bunker capacities
(run 1 - 6), increases in bunker capacities have a greater effect on
efficiency under Policy II than was apparent under Policy I.
lo,
The foregoing results from simple policies indicate that substantial
scope for improvement in policy design exists, and that a combination of the
merits of Policies I and II should be the first step, That is, we require a
policy which employs the maximum bunker discharge rate wherever possible, and
allows for intermediate discharge rates depending on the bunker levels, The
overall aim being to maximise the utilisation of the available belt capacity.
Further, however, it is clear that both of the simple policies used represent
essentially priority policies, Examination of the distribution of coal losses
by coalface resulting from Policies I and II indicate, as expected, that the last
coalface in line (Coalface 1) always suffers the heaviest losses, Such an
uneven distribution of delays between coalfaces will, in fact, always result
from any policy where the limiting factor of the situation, in this primarily the
conveyor belt capacity, is superimposed after the desired discharge rate has been
calculated, It follows, therefore, that any policy dqsign for bunker discharge
rates should also include an allocation of the available belt capacity betveen
the bunkers,
Bunker discharge rates based on allocated belt capacity ~ definition and results
icy ITT
Thie policy was designed on the basis of the points discussed in the last
section, Here it is assumed that if a bunker is full it will be discharged at
its maximum rate, Any residual belt capacity will then be allocated between
unfilled bunkers, in proportion to the ratio of their individual levels to the
sum of levels in the unfilled bunkers, The following steps are hence involved:
(i)_Determine how many bunkers are full
noe = "Yo pe (cn
where NOBF = number of bunkers full
BUNI = level of bunker 1
BUN2 = level of bunker 2
BUN3 = level of bunker 3
BCAP'= bunker capacity ( same for all bunkers)
a
12,
Uy
198
Gi)_Decarmine) bale capacity:to be allocaced The foregoing policy was applied to a similar set of simulation runs
as outlined in Figure 4 and the results of these applications are shown under
ch = BELT- [oBrepyax) wear itty the heading of Policy III in Figure 5, It can be seen that Policy III gives
an improvement in overall efficiency, over Policies I and II, on all runs.
where CL = belt capacity renaining to be allocated = BELT if ell bunkers not full This is due to a coubination of the facts that better use is now being made of
BELT = belt capacity conveyor belt utilisation and also that discharge is not now on a priority basis,
DMAX = maximum beakee discharge rate In particular, the effect of increasing the bunker capacity at low belt
capacities is much more marked, For example, comparison of runs 1 and 3 for
(iii) Determine sum of levels in unfilled bunkers (SLIUB) Policy II in Figure 4, show a 4,8% improvement in,efficiency, whilst comparisons
SLIUB = suMB - [ronr*Bcap} eeees Gy) of runs 1 and 3 for Policy III in Figure 5 show an 8,3% improvement in efficiency,
In other words, the improved belt utilisation, vhich results from the improved
vhere SUB © sum of bunker levels control policy, can be interpreted as @ saving in bunker capacity to achieve a
given efficiency, As in Policy I results run 8 for Policy IIX confirms that the
(iz) Determine individual bunker discharge rates maximum efficiency of the system, under the coalface conditions simulated, is 91,82,
+ DOIA = cLipQouax, BEE cL BUN ,RCAP) @) DEFINITION OF EXPERIMENT OVERALL SYSTEY EFFICIENCY
s} Total | Maximum Capacity | Bunker Discharge | Bunker Discharge
vhere DDIA = desired discharge rate Bunker 1 (A) capes ieee Sober Be tintin mee
F ity Jesch Bonker | (ons) | Allocation” | Allocation”
This will always ensure that the bunker discharges at its maximum rate if (Tons/ | (Tons /Hour) Policy ) Policy )
full, otherwise it will receive a portion of the belt capacity left. “=
ADISL = cure @2, pora,0,poza) was (wi) BLOCK 1] RUN 1 | 2000 1000 500 73.2 73,3
: storacE | RUN 2 | 2000 700 500 7.5 733
vhere ALIS] = Actual discharge rate Bunker 1 BUNKER | RUN 3 | 2000 1000 1000 81.5 81.6
system | RUN 4 | 2000 1000 1200 84.6 84.9
This second stage is necessary because if all bunkers are full and total *
belt capacity is less than the sum of all maximum discharge rates, then negative BLOG 2.) R0H-1 | 2500 1000 150 78.9 79.4
capacity renaining (CL), and hence negative DDIA, will result, In this case the SURGE, (RON 2 (|) 2500; | -to09 #00 88.0 8.7
belt capacity is apportioned equally betveen the bunkers, BUNKER | RUN 3 | 3500 1000 150 84.2 84.5
systex | RUN 4 | 3500 1000 500 91.8 91.8
The sctual discharge rates for the other bunkers are determined by similar equations,
Figure 5 Overall System Efficiences for
* This is a DYSNAP function the interpretation of which fa: Bunker Discharge Policies III & IV
If X= CLIP (A,B,C,D)
Then X= A if CeD
X= Bifc<D,
13,
Of special interest in the results from Policy III, is that associated
with run 2, which shows the effect of reducing the maximum discharge rate of the
bunkers and emphasises the importance of this parameter, Even though the
maxinua discharge rate used for each bunker of 700 tons per hour (2100 tons/hour
in total) is in excess of the total belt capacity (2000 tons/hour) for this run,
a reduction in efficiency is shown over run 1, where the maximum bunker discharge
rate was 1000 tons/hour, This is due to the fact that circumstances can arise
in the application of Policy III, where the belt capacity allocated to a bunker
is in excess of its maximum discharge rate and is, consequently, terminated at
this value as directed in equation (v), This results in a loss of belt
utilisation, Hence Policy III can only be used where maximum bunker discharge
rates somewhat in excess of basic bunker input rates are specified. Since this
arrangement is not common in practice, provision must be made in policy design
This can be achieved by
modifying Policy ITI to carry out e secondary allocation of any spare belt
for the effect of low maximum bunker discharge rates.
capacity, resulting from such termination, between the other bunkers as outlined
in the next section,
Policy IV
This policy represents an extension of Policy III to overcone problems
associated with low maximum bunker discharge rates, Having set all full bunkers
to discharge at their maximum rates, as in Policy IIT, Policy IV then determines
vhether or not the allocation of the remaining belt capacity to the other bunkers
will result in their maximm discharge tates being exceeded by calculating the
This is referred to as
If the bunker level exceeds this saturation level then the
level of each bunker at which this situation will occur,
the saturation level.
discharge of that bunker is also set to its maximum and any remaining belt
capacity allocated between unsaturated bunkers in proportion to their levels,
‘The following steps are involved over and above those outlined in Policy III,
G) Calculate the bunker saturation level (SLEV)
SLEV = i * SLIUB Pert (wii)
199
1,
(ii) Determine the number of bunkers saturated (NOBS), which will include
those full:
BUN |BUN2 |BUN3-
‘NOBS = INT ey. INT po. Int pn] wees (iii)
(iii) Determine the new belt capacity left to be allocated A
assuming all saturated bunkers will be discharged at DMAK:
NCL = BELT - (NOBS*DMAX) wanes (ix)
(iv) Determlne the sum of levels in unsaturated bunkers (SLINS):
sams = wm [sum,suev) + aan [pune,stev] + mw fauns,suzy)
~ frous*szzv] crane @
@)__Determine individual bunker discharge rates
DDIA = MIN (vax, BM xe) tee (i)
ADIS = as equation (vi) tenes Gii)
The actual discharge rates for other bunkers are determined by
similar equations,
The results of applying this modified policy to the previously
defined runs are again shown in Figure 5, where it can be seen that there
is an improvement in efficiency over Policy III in all runs, except,
of course, run 8. Although these differences and efficiencies are
apparently small, an example of their significance is demonstrated in
Figures 6 and 7, which show how the rate of coal arriving at the mine
surface compares with the belt capacity for Policies III and IV respectively.
Run 2 (Figure 5) now achieves the same efficiency as Run 1, confirming that
Policy IV is independent of the maximum bunker discharge rate. In fact,
Policy IV totally maximises the utilisation of the available belt capacity
and hence provides a method of control which cannot be improved for any
given set of system parameters. In addition it will be noted that runs 1
and 2 in Figure 5 are now identical, indicating that Policy IV is also
independent of the maximum bunker discharge rate.
Conveyor
Belt
Loading
Rate
@ors/iour
Conveyor
Belt
(Tons/Hour)
4000.,
35004
3000.1
a3
ES eS
4
Cae Tram os
Rates of Coal Flow at different points on the
Conveyor Belt for Bunker Discharge Policy IIT
(First Belt Allocation Policy)
4000.
Figure 6
Total Conveyor Belt Capacity
wo >
i At the Surface
(h Between Bunkers 1 & 2
Between Bunkers 2 & 3
1m —>
(hours)
Total Conveyor Belt Capacity
Lil WA
ey
Rates of Coal Flow at different points on the
Conveyor Belt for Bunker Discharge Policy IV
(Second belt allocation policy)
Figure 7
KR At the Surface
Between Bunkers 1 & 2
Between Bunkers 2 & 3
TRE
(Hours)
00
15.
QUANTIFYING THE BENEFITS OF ALTERNATIVE BUNKER DISCHARGE
POLICIES
The foregoing results clearly indicate that it is possible to
improve the overall efficiency of a coal clearance system by increasing
the belt and/or bunker capacities and/or by instigating more sophisticated
control over bunker discharge rates; which leads to the interesting
question as to which alternative method should be employed, This is best
answered by considering how much bunker capacity would be necessary
to achieve maximum efficiency for each belt capacity, and bunker discharge
policy used, Such figures were determined by repeating the previous
sirulation runs under the assumption of infinitely large bunker capacities,
and measuring the maximum level achieved in each bunker, These results
are shown in Figure 8,
I Bunker Discharge Control Policies
Belt POLICY I POLICY IT POLICY IIL POLICY IV
Capacity Maximum Bunker | Total Maximum | Total Maximum | Total Maximum
(tons/tr)} Level (fons) | Bunker Level | Bunker Level | Bunker Level
(fons) (fons) (Tons)
2000 7,564 6,150 5,372 5,300
2500 7,564 ij 3,000 2,513 2,390
3500 1,288 1,854 1,346 1,288
Figure 8, Feasible, total maximum bunker levels required
to achieve maximum efficiency (i.e. no coal losses)
for each combination of belt capacity aifd bunker
discharge policy.
In Figure 8 the bunker level quoted is the sum of the maximum levels
achieved in each bunker. All runs, with the exception of those
associated with Policy II, resulted. in the attainment of maximum system
efficiency of 91.8% with no coal face stoppages. Under Policy II
problems occur in determining the maximum bunker levels required. Since,
as explained in an earlier section, the bunker discharge policy itself
interacts with the bunker level. If a large bunker capacity is introduced
the bunker levels rise to use the available capacity and the efficiency falls.
201
16.
The problem cannot be totally overcome without destroying the policy,
but can be partially overcome by fixing the bunker .capacity not at an
infinite value, but at a level, only slightly in excess of the maximum
level anticipated in each run. The results for Policy II in Pigure 6
were determined in this way and, hence, represent approximation to the
total bunker capacity required.
Nevertheless, the results of Figure 8, which are interpreted
graphically (for Policies II, III and Iv) in Figure 9, clearly show?
(i) that for a given bunker discharge policy there is a trade off
between bunker capacity and belt capacity to achieve maximem
efficiency, i.e. at low belt capacity very high bunker capacity
is required which decreases as belt size is increased.
The ultimate extension of this fs that at very high belt capacity
no bunkerage is necessary and that at very low belt capacities infinite
bunkerage is necessary,
(ii) that improved bunker discharge policies can reduce the physical
capacity necessary to achieve maximum efficiency,
Obviously the ultimate criteria of choice between the alternative
methods of achieving maximum efficiency is that of cost, and Figure 10
presents the results of Figure 8, converted to cost terms. Each belt/
bunker combination, has been converted into total capital cost terms by
summing the required belt capacity at £300/tons/hour to the required
bunker capacity of £750/ton, These are average costs/unit of capacity
and are taken as approximately representative of those necessary to
uprate capacity (10).
POLICY III
w
POLICY IV
POLICY If
3.5]
3
2.5)
2
conveyor
belt
capacity
(tons/hour)
Total system bunker capacity (tons)
Figure 9 Variation in conveyor belt capacity with total system bunker capacity to attain maximum
* system efficiency (Policy IT, II and IV)
202
1.
It will be seen that the lowest total cost of each bunker/belt
combination occurs for any given policy by using the largest belt size,
which implies that it should always be more economical to maximise belt
size and minimise bunker capacity, This does, however, ignore the risk
of breakdown associated with such an arrangement, It can be shown that,
in fact, unit conveyor belt costs need to be of the order of four times
bunker costs before smaller belt sizes and larger bunker combinations
become economical on a simple average cost criterion,
By reading across the rows, Figure 10also shows the order of savings
associated with the type of control policy used, which are attributable
to better conveyor belt utilisation and hence, lower bunkerage requirements,
In all cases these will be seen to. be very substantial, As a more moderate
~. example the effect of using Policy III rather than II at a belt capacity of
2500 tons/hour, results in a capital saving of £366,000 (less of course
the cost of control),
BUNKER DISCHARGE POLICY
POLICY I POLICY IT POLICY IIT POLICY 1V
Belt Bunker Cost| Belt Bunker Cost | Belt Bunker Cost|Belt Bunker Cost
Capa- Capa- Capa Capa- Capa~ Capa~ Capa~ Capa~
city city city city city city city. city
Tons/|Tons fm | Tons/| Tons fm | Tons/j Tons im | Tons/} Tons fn
tr Br ir ir
2000 | 7,564] 6.273] 2000 | 6,150) 5.213] 2000 | 5,372 {i 29 2000 | 5,300) |4.575
2500 | 7,564 ||6.423]) 2500 3,000)[3 ,c00| 2500 | 2,513 [2.634] 25¢0 | 2,390) [2.542
3000 | 1,288] [2,014] 3000 { 1,854)[2.440l/ 3000 | 1,346] [2.059]} 3000 } 1,288) (2.
’
Figure 10 Capital costs of each belt/bunker combination
for each bunker discharge policy,
20,
IMPLICATIONS OF RESULTS AND CONCLUSIONS
The report demonstrates that coal clearance systems can be
effectively modelled using system dynamics techniques and that,
since such models incorporate continuous information feedback they
allow a system to be designed in the operational’ control sense as
well as in terms of engineering capacity, In addition to aiding the
development of automatic control rules for bunker discharge rates, models
af the type developed have been shown capable of quantifying the trade off
between improved control and the total physical capacity of any coal
clearance system. The results presented clearly show the relative merits of
increasing system efficiency by improved bunker discharge control,
compared with the alternatives of increasing either conveyor belt or bunker
capacity. This has obvious implications both in the design of new systems
and in improving the efficiency of existing ones.
Using the model, a control policy based solely on bunker levels has
been developed and tested, which is capable of maximising conveyor belt
utilisation, under any range of values of physical parameters in the coal
clearance system described, It will be noted that since such maximisation
is achieved then by definition no further scope for improvement exists,
for any specified combination of parameters, This is particularly
important in the context of using additional information sources in
control rules, For exazple, many other sources of information other than
bunker levels could be specified on which to base buaker discharge control
rules, In particular, it is often postulated that bunker discharge
policies could beneficially be based on coalface information, in addition
to bunker level information; either in the form of current average output
rate or the time to the next planned stoppage, The effect of such
modification to the developed control rules is under investigation in this
and other system configurations but, since conveyor belt utilisation can
be maximised under the existing control rule, this additional information
is likely to be redundant,
It should be emphasised that the model and results described,
represent a situation where automatic continuous scanning of the states of
the system can be carried out, and indicate what is possible if automated
control action could be continuously fed back to the bunker discharge
equipment, This presupposes that continuous adjustment of the equipment
is possible and, in practical cases, the syatem performance will be
proportionately less, depending on the actual bunker discharge increments
which are feasible,
203
al,
ACKNOWLEDGEMENTS
The authorwould like to acknowledge the assistance of J,Zachoval,
R,G,Coyle and A,K,Ratnatunga, for their co-operation in various stages of
development of this research,
BIBLIOGRAPHY
@) Whitworth K, "Advanced Technology Mining", World Coal, 1977.
(2) Chandler K,W,, "On-Line Computer Control Systems for Collieries",
TT Mi Evy May 1974,
(3) . Lawrie S,, “Computers and their Application to Colliery Information and
Control Systems", Mining Technology, December 1976.
(4). Beyon I,, "Computers and Automation", Mining Technology, October 1974,
(5) Tregelles P,G, & Thomas V.M,, "Control Technology", TsId!sEa, March 1976,
(6) Bond W,J, and Dixon G,T,, “Computer Controlled Coal Clearance at Eagvorth
Colliery", Paper presented to I,M.E., February 1977,
(7) Coyle R.G,, "Management System Dynamics", Wiley, 1977.
(8) Ratnatunga A,K,, "Dynamic Simulation Application Package (DYSU\AP) User Manual",
University of Bradford, System Dynamics Research Group, May 1979.
(9) Wolstenholme E,F, "The Incorporation of Discrete Factors in System Dynanics
and Coyle R,G, Models", DYNAMICA, Volume 6, Summer 1980.
(10) Private Communication,
